C24 - Single Equation Models; Single Variables: Truncated and Censored Models; Switching Regression Models; Threshold Regression ModelsReturn

Extreme value theory is the most scientific approach to an inherently difficult problem - predicting the possibility that an extreme event will occur. Broadly speaking, there are two kinds of models for extreme values. The first group of models are models for a distribution of normalized maximum (minimum) of the sequence of independent identically distributed random variables. The second, more modern, group of models are the peaks-over-threshold (POT) models. These are models for all large observations which exceed a threshold. This paper is concentrated on the first type of model. Here, the maximum loss of a bank caused by different technological accidents and natural disasters is treated.

In the article the length of unemployment in the Czech Republic is treated. A nonparametric estimate of its survival function is constructed using maximum likelihood estimation for the groups of men and women and both groups are compared. The positive influence of education on the length of unemployment is illustrated. Data describing the unemployed were gathered by the labour force sample survey organized by the Czech Statistical Office in years 2000-2004.